Calculation of self-diffusion coefficients using MD simulations

[rain]

Dear Dr. Dubbeldam,
To estimate self-diffusion coefficients, I employed RASPA to perform molecular dynamics simulations, followed by calculating the slope of the mean square displacement with respect to the simulation time.
However, the result obtained from the MD simulation, as shown in the plot given in the attachment, shows that there is no strong linear relationship between MSD and time.
I want to know why it obtained such results and how to get an accurate linear relationship between them.
Best regards.

[David Dubbeldam]

Your molecule has not diffused at all. I would first do a test with a small molecule so that you see what a graph looks like for a molecule that diffuses.
The y-axis is the MSD, in Angstrom squared. You can get diffusion results when the molecules diffuse at least through the repeating unit (so unit-cell-width squared).
In this case, MD is probably not capable of computing diffusion because it is too slow in your system.